Methods for detecting a protein-protein interaction can be roughly categorized into two groups. One is a method characterized by using a protein having been separated from living cells. Examples of such a method include surface plasmon resonance, protein mass spectroscopy, and anisotropy measurements. However, these methods have difficulty detecting an interaction in an environment similar to an actual intracellular environment.
Then, as the second method, a method has s been developed, in which a protein-protein interaction is detected using living cells. Typical methods thereof are a yeast two hybrid system which detects a transcriptional activity of a reporter, and modified methods thereof. Besides, another method has been also developed, which utilizes reconstitution of enzymes such as β-galactosidase and dihydrofolate reductase (DHFR).
Nevertheless, these methods have problems that they are incapable of detecting a position where a protein-protein interaction has taken place (positional information on the protein-protein interaction), as well as a period until a protein-protein interaction takes place, a period until the interaction ends, a duration of the interaction, and the like (temporal information on the protein-protein interaction).
Meanwhile, the method for detecting a protein-protein interaction using living cells also includes a method utilizing reconstitution of a fluorescent protein. Nevertheless, once reconstituted, the fluorescent protein does not dissociate. Accordingly, this method has a problem that it is incapable of detecting a period until a protein-protein interaction ends, a duration of the interaction, and the like. Further, there is another problem that a period until a protein-protein interaction takes place and the like cannot be detected because emission of fluorescence requires a certain time after a protein-protein interaction takes place.
Furthermore, there is also a method utilizing a luciferase reconstitution technique. In such a method, a luciferase is reversibly reconstituted and dissociated. However, since the luminescent signal emitted by a reconstituted luciferase is weak, the exposure time has to be a long in order to obtain intracellular positional information, and positional information and temporal information on a protein-protein interaction with high turnover rate cannot be obtained.
Additionally, in the methods utilizing the reconstitution of a fluorescent protein, a luciferase, or the like, a signal can be detected only after such a reconstitution. This also brings about such a problem that it is difficult to trace, for example, both before and after a protein-protein interaction takes place, proteins which are located at different positions by the interaction.
On the other hand, as a method for detecting a protein-protein interaction in living cells, fluorescence resonance energy transfer (FRET) has been developed, which detects energy transfer dependent on a distance between molecules. This method has an advantage of obtaining positional information and temporal information on where and when a protein-protein interaction takes place. Nevertheless, since a positional relation between a donor fluorescent protein and an acceptor fluorescent protein used in the method is important to detect the protein-protein interaction, the method involves a complicated step of investigating the optimization of a linker (spacer) connecting these fluorescent proteins to a detection-target protein, so that such a system has been difficult to construct. Further, it has also been difficult to analyze the result due to cross excitation by which an acceptor fluorescent protein is excited, and to bleed-through in which fluorescence of a donor fluorescent protein bleeds through a filter (absorption filter) set for detecting fluorescence of an acceptor fluorescent protein. Moreover, use of fluorescent proteins of two colors (donor fluorescent protein and acceptor fluorescent protein) also brings about a problem that only limited fluorescent proteins are usable in order to detect information other than a detection-target protein.
Recently, Tobias Meyer et al. have reported a method for detecting a protein-protein interaction by utilizing intracellular localization (translocation) (PTL 1). In this method, one of proteins subjected to interaction detection is fused to a protein that specifically binds to a particular site in a cell, while the other of the proteins subjected to interaction detection is fused to a fluorescent protein or the like. Then, these fusion proteins were expressed in a cell, and the protein-protein interaction is detected on the basis of a signal of the fluorescent protein or the like at the particular site in the cell.
In addition, Nibert et al. have reported a method for detecting a protein-protein interaction, using a fusion protein in which one of proteins subjected to interaction detection is fused to a protein for forming a viral inclusion body, and using, as an indicator, accumulation of the other of the proteins subjected to interaction detection in the viral inclusion body (PTL 2).
However, in these methods for detecting a protein-protein interaction by utilizing intracellular localization, one of proteins subjected to interaction detection is forcibly (artificially) translocated and confined at a particular site in a cell. Accordingly, the detection is impossible at a site where a protein-protein interaction naturally takes place, that is, in an intracellular environment unique to the protein-protein interaction, which brings about a problem that positional information on the protein-protein interaction cannot be obtained, and other similar problems. Moreover, it is also impossible to detect the interaction between proteins localized in a natural state at the same site as the site of the intracellular localization.
Against this problem, Sara Peterson Bjorn et al. have reported a method for detecting a protein-protein interaction (redistribution-trap method), in which proteins are allowed to interact with each other in an intracellular environment where the proteins naturally function, and then the cells are stimulated with a drug or the like to thereby induce aggregate formation from the interacting proteins, the aggregate formation being indicative of the interaction (PTL 3).
However, this method needs to stimulate cells at certain time so that the aggregate formation can be induced, and also needs to remove the drug or the like used for the stimulation to detect the presence or absence of an interaction subsequently after the stimulation. Hence, the method has problems such as being incapable of obtaining temporal information on when the protein-protein interaction takes place, and incapable of detecting a protein-protein interaction that changes (takes place, ends, takes place again, and so forth) for a certain period and at a certain position.